Economic and Social Correlates of Government Control in South Vietnam

■— /-. M -^ t /?. J4ua*j

«4i~t j u j r~Zr —~

ECONOMIC AND SOCIAL CORRELATES OP
GOVERNMENT CONTROL IN SOUTH VIETNAM '

Anthony J. Russo, Jr.

{/

Introduction

The purpose of this study is to explore the relationship
between welfare and resistance in rural South Vietnam using
empirical quantitative data.

When I undertook this study I held the view that the
war in Vietnam was very much a class war — a war between
the rich and the poor, between the peasants and the city-bred.
This is not to say that I felt that ambitions of the men in-
Hanoi were not important, or that communist ideology was not a
factor; I felt, though, that these things were secondary and
that the forces which sustained hostilities and increasing
levels of violence came directly as a result of the enormous
economic and social cleavages within Vietnamese society and
the fact that these cleavages were highly relate to the
usual differences between urban and rural cultures. The
•government in Saigon was comprised of men who were richer,
from urban backgrounds, had formal educations, and were
highly influenced by the French colonial presence (Most had
fought on the side of the French in the war against the Viet
Minh.). The Viet Cong were peasants, had a strong Vietnamese
identity, and had inherited the legacy of the Viet Minh.

*

'

2

Needless to say this view is not a universal one. In
fact, a study by E.J. Mitchell claims to show that quite
the opposite is true — the richer peasants support the
Viet Cong, not the poorer*. Mitchell presents a detailed
statistical model to support his contention, showing a
positive relationship between government control and a

measure of inequality in the sizes of farms. He sees the ,.

results of his study as being something of a challenge to

U.S. policy in Vietnam which he views as giving emphasis to \

economic reform. Actually, this is quite untrue; the U.S.

has paid unprecedented lip service to the role of economic

development and reform, but in practice has done little to

promote it. The reason for this, most likely, is the in-

ability of U.S. policy-makers to agree on the importance of

development and reform**, and their persistence in viewing the

conflict as one in which conventional military strategies

(e.g., search and destroy operations and heavy reliance

upon artillery and air power) are the only ones likely to

provide substantial payoffs. In South Vietnam there has been

*Mitchell, E.J., "Inequality and Insurgency: A Statistical
Study of South Vietnam," World Politics , April, 1968.

**"Land Reform in Vietnam," 20th Report by the Committee on Govern¬
ment Operations , House Report No. 1142 , March 5, 1968. This
report draws attention to differences among U.S. officials which
have impeded progress in agrarian reform (p. 15). Generally, the
report is critical of the lack of progress, but an anti-reform
statement is appended to the report by dissenting members of the
committee who use the Mitchell study to support their position
(p. 19).

- 3 -

no major effort made to promote economic or social reform.

This study finds that geographic variations in rural in¬
come, farm sizes, and religion are highly correlated with
geographic variations in the extent of control by the Saigon
government (Saigon control is relatively high where rural in¬
come is relatively high.). The Mitchell paper is referred to
frequently throughout because the findings shown here are in
substantial conflict with it.

Standard multiple regression techniques comprise the
principal statistical tools employed and two approaches are
taken: one analyses variations in province aggregates of

data (using I960 province boundaries) and the other analyses
variations in data on the hamlet level. Unfortunately, it
was necessary to exclude the Central Highlands from the study
because complete data were not available for that region; the
analyses include the Central lowlands (referred to hereafter
as the Central region) and the areas referred to by the U.S.
Government as III and IV Corps (referred to hereafter as the
Southern region).

\

4

II Province Analysis

There are many ways to view inequality in distributions
of welfare and more than a few indices available for quantifying
the concept. Once we select a welfare measure, or measures,
we can study the differences in the means, or the differences
in the differences (the between-group variance of the within-
group variances). There are 26 provinces in the principal
areas of South Vietnam — the Central and Southern regions
and data are available for each. In this study the
approach is to analyse variations in the Saigon Government's
control from province to province in the context of variations
in measures of welfare and record the degree to which they
are related. The effects of religion and pure regional
differences will also be accounted for. The Hoa Hao religion
is important because the Hoa Hao, whose principal strength
lies in 5 southern provinces (see table 1), exert a degree
of local autonomy not seen in any other part of the country.
Regional differences are important because the Central region
(from and including Binh Thuan province northward to the 17th
parallel) and the Southern region represent different levels of
homogeneity; it is almost, but not quite, as though the two
areas were two different countries. The Central region was
settled long before the Southern region, experienced less

- 5 -

French influence, and has maintained a strong position with
regard to traditional Vietnamese culture, customs, and habits.
The Southern region, on the other hand, was settled more
recently in history when the Vietnamese drove out the Cambodians
(the Southern area was once part of the Khmer Kingdom) and
French influence, later, was more profound with the French
figuring heavily in shaping the economy of the region and its

land tenure characteristics. Religious practices are also

. \

quite different; in the Southern region there are the Hoa Hao

and Cao Dai sects, whereas, in the Central region there are
no Hoa Hao and very few Cao Dai.

The hypotheses investigated are the following:

(1) Saigon control increases (or decreases) ih direct
proportion to increases (or decreases) in measures of
rural welfare.

(2) Saigon control is higher in provinces with substantial
percentages of Hoa Hao population and increases pro¬
portionately as the Hoa Hao population increases.
Relationships between control and welfare differ in the
two Southern and Central regions with the relationship
being stronger in the Central region.

( 3 )

- 6 -
Table 1

PROVINCE DATA MATRIX

I L/Ls HH g D s

Per Capita Ratio of Percent- Gini Index Southern

Income,**** Mean to

Province

Thousands of
Piastres

Subsistenc
Land Hold¬
ing^

An Giang

3.94

10.5

An Xuyen

4.38

14.3

Ba Xuyen

3.90

13.7

Bien Hoa

4.75

4.6

Binh Dinh

2.36

2.3

Binh Duong

4.31

3.9

Binh Thuan

2.90

4.6

Dinh Tuong

6.33

7.6

Khanh Hoa

' 3.34

4.2

Kien Giang

4.40

9.6

Kien Hoa

4.80

4.6

Kien Phong

4.02

10.1

Kien Tuong

5.17

14.3

Long An

4.71

9.4

Long Khanh

5.59

4.8

Ninh Thuan

4.05

5.5

Phong Dinh

3.87

8.3

Phuoc Tuy

6.64

5.1

Phu Yen

2.92

3.1

Quang Nam

2.83

1.5

Quang Ngai-

2.22

2.4

Quang Tri

3.32

1.5

Tay Ninh

3.32

5.7

Thua Thien

3.90

2.7

Vinh Binh

4.53

•"9.3

Vinh Long

5.96

5.9

age Pop¬
ulation in
Hoa Hao
Religion***

of Inequality
in Size of

Land Hold¬
ings**

Dummy

Variable

75

.62

1

0

.44

1

1

.60

1

0

.67

1

0

.38

0

0

.54

1

0

.66

0

0

.49

1

0

.62

0

1

.69

1

0

.57

1

28

.60

1

26

.48

1

0

.42

1

0

.55

1

0

.74

0

10

.42

1

0

.69

1

0

.59

0

0

.40

0

0

.45

0

0

.45

0

0

.41

1

0

.50

0

0

.45

1

20

.44

1

Dc

Central

Dummy

Variable

C '

Percentage
Hamlets
Under Saigon
Control
(1965)*

0

75

0

27

0

24

0

31

1

Q

y

• 0

18

1

32

0

31

1

43

0

22

0

24

0

42

0

45

0

32.

0

25

1

76

0

28

0

52

1

13

1

18

1

IS

1

12

0

22

1

33

0

19

0

64

Regional

Province

Averages

Southern

4.74

8.34

9.5

.53

34

Central

3.09

3.09

0

.53

—

—

28

6a -

^Control data are those used by Mitchell, op. cit.

**g and L/L were calculated from data given in "Report on the Agricultural Census of Vietnam,"
Department of Rural Affairs, Republic of Vietnam, 1961.

***Estimates of Hoa Hao population were taken from "The Religions of Vietnam in Faith and Fact,"
Southeast Asia Religious Project, Fleet Marine Force Pacific/lMAC (Forward), FPO San Francisco

****Stroup, Robert H., "Rural Income and Expenditure Sample Survey: Preliminary Report," United
States Agency for International Development, 1965.

96602

- 7 -

The multiple regression model is as follows*:

(1) C = -11.9 +8.51+ 16.4 D C I - 37 D c + .75 HH

(2.82) (2.50) (-1.55) (5.43)

R 2 = .66 P = 13.0 N = 26

where

\

C = percentage of hamlets under Saigon control (data
are taken from the Mitchell study)

I = estimates of per capita income, 1000's piastres
(VN$) per person

HH = estimated percentage of province population in the
Hoa Hao religion
D c •= 1 for Central region
= 0 for Southern region

Dummy variables are employed to test the hypothesis that the

J

relationship between control (C) and income (I) is stronger
in the Central region**.

For the Central region (D = 1) equation (l) becomes

(2) C = -48.9 + 24.9 I

*The values shown in parentheses beneath the regression
coefficients are t values (the ratio of the regression co¬
efficient to its standard error). The higher the t value the
more statistically significant is the regression coefficient.

**A good explanation of the use of dummy variables in this
fashion is given in the following:

Johnston, J., Econometric Methods, McGraw-Hill Book Co.,

1963, P- 221-228.

- 8 -

For the Southern region (D = 0) equation (1) becomes

(3) C = -11.9 + 8.5 I + .75 HH

Table 2 - Correlation matrix for province data

C

1 l /l s

HH

S

L/L

C

I

1

.40 .28

1 .35

.60

.08

.39

1

.45

.19

.05

.10

1

1

HH

The Hoa Hao term is dropped in equation (2) because
there are no Hoa Hao in the Central region.

In other words, in the Central region, the percentage
of hamlets under Saigon control increases 24.9 percent for
each increase in rural per capita income of 1,000 piastres
(5 to 10 U.S. dollars depending upon which exchange rates
are used). In the Southern region, control increases 8.5 per
cent for each increase in per capita income of 1,000 piastres
and .75 percent for each 1 percent increase in Hoa Hao population.

The three hypotheses stated above are strongly supported
by the statistical analysis thus far. Data from only one
year (1965) have been used, however, control patterns have
not changed markedly over the course of the war. Areas of
strong Viet Cong influence in the early years of the war
remain strongly pro-Viet Cong. Also, as will be shown later
in the hamlet analysis, 1967 control data show about the

same relationship with income.

—2

K is the multiple correlation coefficient squared;
its numerical value is equivalent to the percent of variance
in the dependent variable that is said to be “explained".

In the regression model shown in equation (1), 66$ of the
variance is explained. The remaining 34$ of the variance
which is unexplained may be attributable to error in the data
or variables which have not been included in the model.

There are other measures of rural welfare besides that
of income. Another variable which provides additional ex¬
planatory power in a model similar to equation, (l) is the
ratio of the mean size of a landholding (L) to a “minimum
subsistence" landholding (L ). L Q is the amount of land re-

O O

quired to provide an average Vietnamese family (5.3 persons)
with a daily rice ration of 250 grams per person for one year;
it reflects variations in rice yields across South Vietnam
and ranges from .17 hectares in some provinces in the Mekong
Delta to .40 hectares in Quang Tri — the northernmost
province. L/L o reflects the extent to which farmers are
securely above the minimum subsistence level. When L/l s
is added to the mix of independent variables shown in
equation (1) it is found to be highly significant in the
Central region and to have no effect in the Southern region;

10

we might expect this because L is "contaminated" to a certain
extent in the Southern region by the presence of very large
holdings; this makes interpretation of L uncertain because
it is not clear how it reflects the sizes of family farms.

In the Central region, however, L is a clear reflection of
sizes of family farms because there are virtually no large
holdings.

The results of the regression analysis including L/L s
are as follows*:

(4) C = -33.9 + 9.0 I + .76 HH + 10.9 _L_ - 10.9 _L_ D s + 17.3 D s

1 g 1 s

(4.51) (7.18) (5.43) (-5.13) (1.68)

R 2 = .82 P = 22.5 N = 26

For the Central region (Dg = 0) equation (4) becomes

(5) C = -33.9 + 9.0 I + 10.9 JL

L s

For the Southern region (D = 1) equation (4) becomes

(6) C = -16.6 + 9.0 I + .76 HH

Equation (4) is a significantly better predictor than equation (1)
82 percent of the variance is explained (compared to 66 percent
in the first model) and. the t values as well as the overall F
value are much stronger. The difference between the two models

is used instead of ]£ because it shows in a more straight¬
forward fashion how the L/L s term drops out completely for the
Southern region.

11

is apparent almost entirely in the equations for the
Central region. Note how the equations for the Southern
region — (3) and (6) — are virtually the same, whereas,

the equations for the Central region — (2) and (5) —

show that when L/L o is included it carries a share of ex¬
planatory power previously carried by income alone but it
does a more precise job of it in that a higher multiple
correlation is achieved.

It would make sense, in the first place, to add a land
variable to an income welfare index in the Central region be¬
cause land is more of an issue there; there are relatively
more farmers, their holdings are much smaller and less fertile
and their aspirations for land, naturally, are much greater.

The Stroup Survey gives an occupational breakdown for rural

*

heads of household. In addition to the farmer category
there are merchants, peddlers, fishermen, clerical workers,
laborers (skilled, semi-skilled, and unskilled), etc. In the
Central region 65$ were listed as farmers compared to 53$ in
the Southern region. Also, in the Central region, the average
holding is .67 hectares compared to 1.71 hectares in the
Southern region. The Stroup Survey also showed that when
asked what they would buy with a hypothetical increase in
income of 10,000 piastres per year, 37$ of respondents replied

Op. cit.

12

"buy land" in the Central region compared to 22$ in the
Southern region.

Having seen empirical support for the hypothesis that
Saigon control is correlated with higher income and, in the
Central region, larger family farm$ we ask how this can be
reconciled with Mitchell's finding of a positive correlation
between his measure of inequality and control and his sub¬
sequent conclusion that'the poor support Saigon more. Briefly,
the answer lies in the fact that income and inequality are
positively correlated, not negatively as Mitchell assumed.

In examining the question I have used the more standard
measure of inequality: the Gini index* (see table 1). I
find that the C-ini index (g) is indeed correlated (r = .45>

N = 26) with control (C), however, the strength of the cor¬
relation lies in the Central region. In the Southern region
(N =17) the correlation between C and g is .17 which is
not statistically significant, however, in the Central region
(N = 9) the correlation is .80 which is highly statistically
significant. Thus, although there is a significant association
between the two variables in the overall sense,

*For an explanation of the Gini index see, Russett, Bruce M.,
Trends in World Politics, The MacMillan Co., 1965> p. 1=7-121.

13 -

we can be misled if v/e do not examine the tv/o regions
separately. The strength of the relationship lies in the
Central region and there g is highly correlated with income
(r = .57) and L/L s (r = .94).

It was also found that the inequality index contributes
nothing when put into the regression model shown in equation
(4). The Gini index and the coefficient of variation were
both tried in equation (4) and each turned out to be
insignificant while coefficients and t values of the other
variables remained essentially unchanged.

The analysis up to this point has been based upon
patterns in variables which are provincial aggregates. In
the next phase of the study we look at the situation at the
hamlet level; as will be seen the results from the hamlet
analysis are very similar to the provincial aggregate analysis
'.in that income is a strong explanatory variable, and when
combined with the regional dummy variable shows very similar
patterns.

- 14 -

III Hamlet Analysis *

In this phase of the analysis we investigate the relation¬
ship between the degree of Saigon government control and per
capita income at the hamlet level for a random sample o.f
94 hamlets. Hamlet data came from two sources: (1) the extent
of Saigon control is represented by indices taken from the U.S.
Government Hamlet Evaluation System (HES) reports of December,

1967, and (2) per capita income data are taken from information

\

gathered in an income and expenditure survey conducted in the
rural areas in 1964**. The HES data are subjective security
ratings, supplied by local American advisors, on a scale
v/hich varies from 0 to 5 (0 means total uncontested Viet Cong
control and 5 means total uncontested Saigon control); the
data are compiled monthly for virtually all hamlets (approximately
13,000)'in South Vietnam.

We are not concerned with the absolute levels; of the
hamlet ratings but rather their variation from hamlet to
hamlet. Therefore, any bias in the ratings (positive or negative)

*Hamlets are the smallest jurisdictional areas in South Vietnam.
Usually several hamlets make up one village and several villages
make up one district. Provinces are composed of three to five
districts.

**Stroup, Robert H., "Rural Income and Expenditure Sample Survey:
Preliminary Report," United States Agency for International
Development, 1965. The author is indebted to Professor Stroup
for making available the entire set of raw data gathered in the
survey (a total of 2,910 quite detailed household interviev/s).

- 15 -

is not likely to have substantial effects on the relationships
uncovered so long as it is consistently high (or low); random
errors in the security indices will simply make the relation¬
ships appear to be a little less strong (in a probabilistic
sense) than they actually are.

The set of hamlets sampled for the economic survey in
1964 are a biased sample because interviewers were unable to
gather data in areas under strong Viet Cong control. Strictly
speaking, then, the relationships we uncover in the hamlet
analysis apply to that hamlet population not strongly controlled
by the Viet Cong in 1964. This is a meaningful population
about which one can draw inferences, but the more interesting
question is: To what extent is it possible to generalize the
findings to all hamlets, i.e., the population of hamlets that
would have been represented by a sample not biased by security
conditions in 1964? There are no statistical tools, that I
know of, that can provide an unequivocal answer. My feeling,
however, is that the security bias in the sample does not
seriously affect the relationships we find between Saigon
control and per capita income; in fact, I feel it likely that
the security bias weakens the relationships much the same
as the correlation between two variables with a bivariate
normal distribution tends to be diminished when the sample

' \

- 16 -

points with the highest (or lowest) values of one of the
variables are eliminated.

Unfortunately specific data were not available on hamlet
religion or land fertility, therefore, it is not possible to
use the same variables used in the province analysis.

The regression model is specified exactly like equation (1)
in the province analysis, leaving out the Hoa Hao term for
which we have no data. The results are as follows:

*

(7) S = 1.62 + .23 X + .49 XD q - 1.44 D c

(3.25) (2.41) (-1.98)

R 2 = .25 P ='10.1 N = 94

where *

S’ = hamlet security rating (December, 1967)

X = 1964 hamlet per capita income, 1000's piastres
per person

For the Central region (D q =1):

(8) S = .18 + .72 X

For the Southern region (D c = 0):

(9) S = 1.62 + .23 X

In the Central region the hamlet security rating increases
by .72 for each increase of 1,000 piastres in per capita income.
In the Southern region the relationship is weaker; the hamlet

- 17 -

security rating increases by .23 for each increase of 1,000
piastres in per capita income.

The important thing to note here is the similarity be¬
tween equation (7) and equation (1); in both the province
model and the hamlet model the income effect (in terms of the
magnitudes of the regression coefficients) in the Central
region is about three times what it is in the Southern region.
This is illustrated more vividly by showing both equations

(I) and (7) in non-dimensional form (i.e., expressing each
variable in terms of its standard deviation).

Province Model

(10) U c = -.66 + .54 Uj + 1.05 - 2.05 D c + .68 U RH

Hamlet Model

(II) U s = 1.35 + .38 U x + .80 U X D C - 1.2 D Q

The two models show a fair degree of similarity. The co¬
efficients in the hamlet model are slightly weaker than in
the province model and the fixed term is positive, whereas,
in the province model it is negative. It is possible, and
likely, that the security bias, discussed above, is responsible
for the difference. A weakening of the relationship would
tend to make the fixed term have a higher value, and in this

case it does.

18 -

IV. General Comments and a Discussion of the Findings

The main conclusion of the study is that economic and social
factors have been predominantly important in the struggle in
South Vietnam. The variance in control patterns is explained
extraordinarily well without taking into account military or
strategic considerations and the basic relationship between
control and income holds up whether we look at the situation in
1965 (as in the province analysis) or in 1967 (as in the hamlet
analysis)* Certainly there is error in the data and much of
the unexplained variance has to be attributed to it. That
leaves little variance unaccounted for and doesn't say much
for the effectiveness of the vast amount of military resources
that have been expended over the years.

We have not been concerned with the long-run underlying
causes of the conflict in Vietnam, nor have we considered factors
which precipitated hostilities. We have simply been concerned
with the question of which areas are relatively weaker or
stronger in support for one side or the other and the factors
which explain the variance in intensity of support. There is
little doubt that opposition to the various Saigon regimes

^Actually, there has been little change in the geographical
pattern of Viet Cong control throughout the course of the war,
although there may have been small local fluctuations, or vari¬
ations in overall intensity such as the period of time immediately
after the overthrow of the Diem regime in 1963> and after the
1968 Tet offensive. In both cases, the level of Viet Cong control
seemed to use uniformly throughout the country.

19 -

manifests strong '.'class" overtones; class is placed in quotation
marks because we do not see homogeneous groups either wholly for
or against the Saigon government; we see, rather, a continuous
spectrum ranging from poorer to richer areas.

The empirical relationships developed in this study probably
reflect, also, the "cultural distance" between the urban, western
-oriented Saigon regime and the rural, agricultural, and more
traditional peasant. In contrast, however, one is hard put to
find much of a "distance" between the Viet Cong and the peasants.
And this does not mean necessarily that the Viet Cong are loved.
The lack of cultural distance between the Viet Cong cadre and
the peasants means, for example, that he has the same rough
peasant features and wears the same faded, worn-out clothes as
the latter. Y/hen he talks to them, whether they like what he
says or not, they understand him and they understand what their
role is with regard to the system he represents. They know
how their behavior relates to the powers of rewards and punish¬
ments he possesses; he is capable of "honeyed talk", as it is
often described by the peasants, or ruthless terrorism, and he
usually has a specific reason for the use of either. On the
other hand, the cadre who represents the Saigon government in
the hamlet presents an entirely different image; his clean
store-bought clothes and his smooth facial features betray his

20 -

/\) O

class origins and, not always, but often, it is obvious that he is
uncomfortable in his surroundings and just doesn't like the
idea of having to be -there. Just as the peasant doesn't
necessarily love the Viet Cong cadre, he doesn't necessarily
hate the city boy who represents Saigon. He may envy him,
admire his status and wish that he could be like him, but there
are two problems: (l) the peasant cannot relate to him on an
emotional level, nor (2) can he see how his behavior relates
to the authority pattern represented by the city boy; the
consequence is that the peasant sees no role for himself in the
system represented- by the Saigon government. The Allies may
dispense artillery and napalm, or economic aid and medical
care, but the peasant sees no correlation between either one
and his behavior. His choices are limited; the Viet Cong
provide a role for him, but all too often the Saigon govern¬
ment provides indifference and uncertainty.

Since these conclusions are totally in conflict with
Mitchell's-findings it is useful to examine his work more
closely.

One of his most prominent mistakes was to equate his
measure of inequality with poverty* This brings up a very
important point; measures of inequality such as the one

*There are numerous other mistakes in his paper; principally,
the regression model he reports contains terms which are not
statistically significant although he reports them as being
highly significant.

21

\

Mitchell used (and the Gini coefficient too for that matter)
may be very useful if we are examining one group and want an
index to summarize the degree of inequality. If we are com¬
paring groups, however, we have to be careful. One group may
have a low inequality index, but everyone could be equally poor.
Another group may have a high degree of inequality, but the
least well off in this group could still have an adequate
living standard. For this reason we have to have absolute
measures of welfare such as per capita income. (Although
that too has its shortcomings, they are of a second-order
nature compared to what is being discussed here.) This is
precisely the trap into which Mitchell - falls. Provinces
where inequality (as measured by a single index) is highest
are often the richest. We have shown in this study that the
strength of Mitchell's relationship between Saigon control
and inequality is positively correlated with per capita income
and when per capita income and the Gini index (as well as the
coefficient of variation) are put into the same regression
equation, with Saigon control as the dependent variable,
inequality is not significant.

After Mitchell reaches the conclusion that the rich,
peasants have spearheaded the resistance in Vietnam, he asks:

22 -

"In terms of human behavior, how is one to explain these
findings? Part of the explanation may lie in the frequently
mentioned docility and low aspirations of poorer peasants."*

There is no doubt some truth in the general hypothesis that
poorer peasants tend to be conservative and have low aspira¬
tions; however, one h.as to go much farther than that. If we
simply stop there we miss the main point entirely and fail
to understand much about the dynamics of Third World revolutions.

' The important point is that Mitchell stopped short of consider¬
ation of changes in aspirations brought on by agit/prop activ¬
ities by the revolutionaries. When poverty exists in a
society it becomes an issue at some point among certain elements
of the elite, and at the point when members of’ this elite be¬
come revolutionaries a new process gets underway — the process
of agitation and propaganda and the business' of the revolutionary
agit/prop is to change aspirations in the poverty sector in
order to recruit support for the cause of change. The revolu¬
tionary elite, often from the middle and upper economic classes,
are well aware of the docility of poorer peasants and see this
as one of the major barriers to be overcome in the initial
stages of a movement. Thus it is misleading if we consider

*0p. Cit.

0

- 23 -

the poor peasant in an isolated way and ignore his inter¬
action with the elite. Where there is no tradition of re¬
volutionary agitation, the gap between aspirations and in¬
come, i.e., relative deprivation, may increase as we consider
higher and higher income levels. But this is only one kind
of relative deprivation. These aspirations are a function
of standard of living.. The aspirational component due to
agit/prop activities is quite different and, in the end, we
have to consider the combination of both. And there is still
the question of which side the individual sees as being most
likely to fulfill his aspirations. Overall aspirations are
more likely to be solely a function of standard of living in
cases where there is no tradition of revolutionary agitation,
or where it has not yet grown to any significant proportions.
In Vietnam it is well known that the Viet Cong emphasize the
poor in their appeals for support and to a great extent they
have a monopoly on access to the peasant. The Saigon regimes
have suffered from a mix of maladies, one being a patronizing
and contemputuous attitude toward peasants; this has precluded
the formation of a working mass-elite relationship with the
consequence that the authority structure is ill-defined; a
functional role for rural elements does not really exist.

r

- 24 -

Saigon policy reflects their attitudes toward peasants; an
illustration of this is seen in a statement by an ex-Vietr Cong
who had been a Party member in Kien Giang province and, for
years, had worked as an NLP political cadre:*

"It would be hard to work for the dovernment
as a political worker because it requires degrees
and diplomas. I think I am capable of doing psycho¬
logical warfare, but I don't have the required educa¬
tional background."

The Saigon government in the various forms it has taken since
the departure of the French, has been a system in which parti¬
cipants are more conscious of upward loyalty than they are to
responding to pressures from below. There is very little real
viable contact between mass and elite. The Vietnamese peasant
can perceive little realtionship (other than a random one) \ aj ~ l !

between his behavior and the pattern of rewards and punish¬
ments meted out by representatives of the central government.

On the other hand, a hallmark of the Viet Cong's relationship
with the peasant has been a highly organized consistency shaping
a pattern of persuation and coercion in which the latter has no

*RAND CORPORATION INTERVIEW AD-543. The Rand Corporation series
of interviews with’ ex-Viet 'Cong has been made available for
general use by serious scholars.

\

l

- 25 -

problem comprehending his role. In many ways the Viet Cong's
exercise of power can be said to be more congruent with peasant

I

modes of behavior than that of the Saigon government. The lack of
"Vietnameseness" brought on by a Western orientation of the
governmental elite and the presence of the Americans makes
difficult the development of an authority pattern congruent
with traditional ones. The central authority is, thus, caught
in a situation where it is unable to compete with the Viet Cong f
for the resources of its citizens, nor, through political action,
interdict the flow of resources from the villagers to the
Viet Cong.

Thus, there are two components to the overall level of
aspirations amongst the peasants: the increased level which
comes as a result of revolutionary agit/prop activities, and
that component whose intensity is roughly a function of the
peasant's position on the standard-of-living scale. The process
is a dynamic one in which the respective magnitudes of the two *
components change with time. It is reasonable to conjecture that
the question of which side the peasant sees as fulfilling his
aspirations is related to the comparative magnitudes of the two
aspirations components. If there are poor peasants with high
aspirations, the major component is likely to be the result of
agit/prop activity and they are^likely to believe that their

26 -

aspirations will be realized by support and cooperation with
the revolutionaries. Where the other aspirations component
predominates it is likely that the status quo is the prefer¬
able alternative. • .

In summary, the results of this study are in conflict
with earlier work on the same subject which supported an
anti-reform position as being the best strategy for the Saigon
government. This study finds that, in the Southern region of

v

South Vietnam, Saigon control is highest in areas where rural
income is highest and where local autonomy, as in the Hoa Hao
areas, is greatest. In the Central region, Saigon control
is highest where income is highest, where the size of the
average peasant's farm is greatest, and where land is most
fertile. Analyses at tw6 levels of aggregation (province and
hamlet) were found to give similar results, and, in the pro¬
vince analysis, the socio-economic variables employed left
so little unexplained variance in control that one wonders
whether or not the military effort has had any relevance. The
roots of the conflict appear to have been overwhelmingly poli¬
tical and socio-economic in nature.

/

%»

C°lj

%